Um algoritmo genético aplicado ao problema de empacotamento de bins tridimensionais

Abstract

The objective of the present work is to present one technique to solver combinatorial optimization problems, that involves permutation, in an acceptable computational time based on genetic algorithm. This technique partitions the set of viable solutions in small regions, distinct between itself, in order to diversify the local search. We apply the technique to the Three-Dimensional Bin Packing Problem. The problem is NP-hard in the strong sense and extremely difficult to be solved in the practice. Extensive computational experiments are reported for instances with up to 90 items, and the results are compared with those obtained from the literature.